Rebuilding Infrastructure for AI: Storage, Scale and the Rise of Neo Clouds
As AI adoption moves from experimentation to production, organisations are rethinking how infrastructure is designed. Michael Reid speaks with Renen Hallak, Founder and CEO of VAST Data, about why traditional infrastructure models are struggling to support modern AI workloads and how software-defined infrastructure is changing the industry.
The discussion explores the rise of neo clouds, the growing importance of data, and why AI infrastructure requires a fundamentally different approach to storage, compute, and scalability.
Why does AI require a different infrastructure model?
AI workloads place far greater demands on infrastructure than traditional enterprise applications. Rather than simply adding more storage or compute, organisations need platforms that can move, process, and manage massive volumes of data efficiently. Renen explains why software architecture has become just as important as the underlying hardware.
The rise of neo clouds
The conversation explores how a new generation of cloud providers is challenging traditional hyperscaler models. Instead of adapting legacy infrastructure, neo clouds are being designed specifically for AI workloads, with greater flexibility, scalability, and performance. Renen notes that VAST is already working with approximately 100 neo cloud providers globally, highlighting how quickly this market is evolving.
Can traditional storage keep up with AI?
Traditional storage architectures were never designed for modern AI workloads. VAST discusses how software-defined infrastructure removes many of these limitations, with average 3:1 data reduction and potential 10–12× infrastructure efficiency gains helping organisations reduce complexity while improving performance.
Looking ahead: Enterprise AI is still in its early stages
Although AI investment continues to grow rapidly, the discussion suggests most enterprises are still in the early stages of production deployment. Future success will depend on building infrastructure that combines Compute. Network. Storage. into a flexible, software-defined platform capable of supporting AI at scale.
